Updated CMS application physics list settings
The CMS application settings have been updated:
-
Updated region list for the activation of the 2 physics list:
- ~ 85-90 % of the steps are done in the "EcalRegion" that contains only 25 out of the 425 material-cuts couples
- the physics list, that contains EM models with "sampling-table" based final state sampling has been set to be used only in this "EcalRegion"
- the physics list, that contains EM models with "rejection" based final state sampling has been set to be used in all other regions
- this minimises the memory usage (by the sampling tables) while maximises the potential gain
- also, this will make the initialisation fast (as without using any sampling tables)
-
The default EM physics setting was changed to be closer to the one used by the CMS experiment:
- in their Geant4 simulation, the CMS experiment uses the EM-opt1 Geant4 EM physics constructor everywhere except the hadronic calorimeter region "HcalRegion" where the EM-opt0 is used
- the main difference in the 2 constructors is that the MSC model in opt1 has a less strict step limit parameter (good for crystal calorimeters) while opt0 contains the default, more strict MSC step limit parameter value (good for sampling calorimeters)
- therefore, opt1 is a faster simulation with longer steps for changed particles compared to opt0
- in the current settings, the GeantV/Geant4 CMS applications has opt0 settings everywhere
- this has been changed by adding to MSC models to the MSC processes:
- one being active only in the "HcalRegion" with a range-factor of 0.06 (opt0)
- one being active everywhere except the "HcalRegion" with a range-factor of 0.2 (opt1)
- this will bring the CMS applications (both GeantV and Geant4) closer to the real CMS simulation settings/steppings
- also demonstrates the feature of having multiple models in a given process being active in different region of the detectors i.e. the possibility of combining different physics description per-detector region within a single physics list (beyond the feature of having different physics list per-detector region)
Note, that warnings are reported by the GeantV physics framework but everything is correct including these warnings!
Edited by Mihaly Novak